Benchmarking large multimodal models for ophthalmic visual question answering with OphthalWeChat
This study presents the first bilingual VQA benchmark for ophthalmology, distinguished by its real-world context and inclusion of multiple examinations per patient. The dataset enables quantitative evaluation of VLMs, supporting the development of accurate and specialized AI systems for eye care.
Authors
- Mingguang He (ORCID: https://orcid.org/0000-0002-6912-2810)
- Xia Gong
- Danli Shi (ORCID: https://orcid.org/0000-0001-6094-137X)
- J. Yang (ORCID: https://orcid.org/0000-0002-3624-3659)
- Meng Yuan (ORCID: https://orcid.org/0000-0002-0875-0127)
- Xiaolan Chen (ORCID: https://orcid.org/0000-0002-4765-6096)
- Zheng Yalin (ORCID: https://orcid.org/0009-0008-3617-2732)
- Bingjie Yan (ORCID: https://orcid.org/0000-0002-8810-9689)
- Pusheng Xu
- Weiyi Zhang (ORCID: https://orcid.org/0009-0008-2780-9121)
Institutions
- Hong Kong Polytechnic University (HK)
- University of Liverpool (GB)
- École Polytechnique Fédérale de Lausanne (CH)
- Wuhan Institute of Technology (CN)
Publication Details
- Journal
- Advances in Ophthalmology Practice and Research
- Published
- 2025-10-30
- DOI
- https://doi.org/10.1016/j.aopr.2025.10.006
- Citations
- 1
- Primary Topic
- Multimodal Machine Learning Applications
- Type
- article
- Field-Weighted Citation Impact
- 1.28